J. Phys. Chem. 1987, 91, 5155-5158
5155
Solvent structural features are now incorporated and gives in this simple form two new parameters for each solvent, namely the correlation length and the short-range dielectric constant. Ways of extracting these quantities from experimental data and of associating them with microscopic models are available in several recent reports. The data in ref 18 would be in line with solvent nonlocality by the following consideration, which has to be rather qualitative in view of our lacking precise structural information for most of the solvents. As a first crude approach, the aprotic solvents can be regarded as structurally correlated broadly as the parameter e;' - es-'. All the points on the hvmaX/(e;'- e ; ] ) plot would then be shifted toward smaller t;l - e;', and more so the larger 1;e - €,-I. This would give a larger slope and a smaller intercept,
perhaps better in line with the expected small structural differences between the two oxidation states of ruthenium. To bring water closer to the other solvents, a weaker polarization correlation around the solute ions than for the other solvents is needed. There is no direct evidence for such an effect, but large ions sometimes do exert a "geometric" structure-breaking effect on the water hydrogen-bond network.30 Also, in a previous analysis of interionic interactions around large alicyclic ions in water, by means of nonlocal dielectic theory, these ions were found to give much smaller correlation lengths than those extracted from solute aliphatic ions.20,26 In conclusion, two solvent effects can be suggested as possible origins of the different behavior of water and polar aprotic solvents, namely solute hydrophobicity or solvent nonlocality combined with solvent "structure breaking". These two effects should be distinguishable in principle. Hydrophobic effects would thus lead to a tighter hydrogen-bond structure, reflected in increased librational and molecular deformational vibrational frequencies and decreased 0 - H stretching frequencies and dielectric relaxation times.25 Nonlocality and structure breaking would lead to the opposite effects.
(29) (a) Kjaer, A. M.; Ulstrup, J. Inorg. Chem. 1986, 25,644. (b) Kjaer, A. M.; Ulstrup, .I. Inorg. Chem., in press.
(30) Buslaeva, M. N.; Samoilov, 0. Ya. In The Chemical Physics of Solvation. Parr A . Theory of Solvation; Dogonadze, R. R., KBlmBn, E., Kornyshev, A. A., Ulstrup, J., Eds.; Elsevier: Amsterdam, 1985; pp 391-414.
almost equally simple equation
7,20926328b929
Gas-Phase Chemistry of NF(a' A): Quenching Rate Constants E. Quiiiones, J. Habdas, and D. W. Setser* Department of Chemistry, Kansas State University, Manhattan, Kansas 66506 .(Received: May 29, 1987)
The reaction of excess F atoms with HN, in a halocarbon-coated flow reactor has been used to generate and study the quenching reactions of NF(a'A,v'=O) at 300 K. Fifteen reagents, selected to represent a variety of chemical interactions, were tested in order to provide a survey of quenching rate constants that could be compared to O,(a'A,) and NH(a'A). A range of rate constants was found which varied from -1.0 X IO-'] 6171, s-' for N(CH,), to -1.0 X (3117, s-I for N2, H2, and NO. The NF(a'A) molecule is much less reactive than NH(a'A), and it seems to more closely resemble O,(alA,). However, the quenching rate constants for NF(a'A) generally are larger than for 02(a1A) and show a greater variation from one reagent to another. The self-quenching rate constant for NF(alA) is (2.2 k 1.2) X 10-7 cm3s-I with energy pooling to give NF(b'Z+) being a small component of the total self-quenching reaction.
Introduction The metastable NF(a'A;l.42 eV) molecule is one of the few electronically excited states that can be generated in high concentration and with high efficiency by gas-phase chemical react i ~ n . I - ~In order to utilize the NF(a'A) as an energy storage system, it is necessary to know the quenching rates by other molecules, with itself, and with ground-state NF(X%). We have initiated a program to characterize the quenching and reactive properties of NF(aIA), and we now report quenching rate constants for several types of molecules. With the aid of this survey, comparison can be made to the isoelectronic molecule 02(a'A,;0.98 eV), which is generally unreactive and quenches by electronic(1) Habdas, J.; Wategaonkar, s.;Setser, D. W. J . Phys. Chem. 1987, 91, 451. (2) Pritt, A. T., Jr.; Patel, D.; Coombe, R. W. Int. J . Chem. Kinet. 1984, 16, 971. (3) (a) Cheah, C. T.; Clyne, M. A. A,; Whitefield, P. D. J . Chem. SOC., Faraday Trans. 2 1980, 76.71 1. (b) Cheah, C. T.; Clyne, M. A. A. J . Chem. SOC., Faraday Trans. 2 1980, 76, 1543. (c) Cheah, C. T.;Clyne, M. A. A . J . Photochem. 1981, 15, 21. (4) Malins, R. J.; Setser, D. W. J. Phys. Chem. 1981, 85, 1342.
0022-3654/87/2091-5155$01.50/0
to-vibrational (E-V) energy transfer, and to the more reactive NH(a1A;1.56 eV) and O('D;1.97 eV) cases. Ultimately, such comparisons will elucidate the chemistry of these excited states. The experimental method is based upon the generation of NF(a) by the F H N , reaction system in a flow reactor ( 10 m SKI flow velocity): F + HN3 H F N3 (1)
+
N
F
+ N3
-
+
+
NF(alA)
+ N2
(2)
Most experiments were done with excess [F] in order to maximize [NF(a)]. However, reaction 2 is sufficiently rapid that NF(a) can be generated and studied with [F], [HN,],. With excess [F], the HN, is converted to NF(a,u'=O) with an efficiency greater than 0.8.' The NF(a,v'=l) population is insignificant. The quenching studies were done by adding reagents to the downstream part of a flow reactor after reactions 1 and 2 were completed. The [HN,] concentration was limited to
